The threshold interpolation conjecture for consecutive patterns in random permutations

About 3 years old · traced to

Let σn,m\boldsymbol{\sigma}_{n,m} be a uniformly random permutation of [n][n] with exactly mm inversions. For a permutation ρ\rho, write ρ\rho for its complement by replacing each entry ii with ∣ρ∣+1−i|\rho|+1-i, and let σn,m\boldsymbol{\sigma}_{n,m} contain ρ\rho mean that ρ\rho occurs as a consecutive pattern. For a permutation ρ\rho, let inv⁡(ρ)\operatorname{inv}(\rho) denote its number of inversions. Let π\pi be any consecutive permutation pattern, and set

s=inv⁡(π),s′=inv⁡(π‾).s=\operatorname{inv}(\pi),\qquad s'=\operatorname{inv}(\overline{\pi}).

The threshold interpolation conjecture. If n1−1/s≪mn^{1-1/s}\ll m and (n2)−m≫n1−1/s′\binom{n}{2}-m\gg n^{1-1/s'}, then

lim⁡n→∞P[σn,m contains π]=1.\lim_{n\to\infty}\mathbb{P}\big[\boldsymbol{\sigma}_{n,m}\text{ contains }\pi\big]=1.

The preceding theorem establishes the appearance and disappearance thresholds separately, while this conjecture asserts almost-sure presence throughout the interval between them. The paper notes that its random-composition methods do not establish this intermediate regime.

References

Primary source

David Bevan and Dan Threlfall, “Thresholds for patterns in random permutations with a given number of inversions”, arXiv:2312.01182 (2024).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.